A proposal for an AI-governance baseline. Enter the access key you were provided.
Grace Hill has put AI to work where it matters most — inside fair-housing, compliance, and resident-experience workflows. That's exactly why it deserves governance that can stand in front of an auditor, a regulator, or a buyer. We're proposing a fixed-scope baseline that inventories your AI estate, scores it honestly across three lenses, and hands you a prioritized path — in weeks, not quarters.
Your platform embeds generative AI into policy, training, and compliance content that operates across all fifty states' housing law. Regulators have been explicit: HUD's May 2024 guidance signals that the Fair Housing Act can apply to AI used in housing-related functions — and that liability may extend to technology providers, not just operators (HUD PR 24-098). At the same time, boards and private-equity owners increasingly treat documented AI governance as part of exit- and audit-readiness.
To be clear about what we are not saying: nothing in our preliminary review suggests your AI behaves improperly. What we observe — from a review of public materials only (June 2026) — is that we could not locate a published model disclosure, accuracy basis, or human-oversight description for AI features answering compliance-sensitive questions. The remedy for a control gap is straightforward and fast: document, add the right human-in-the-loop gates, and review — which is precisely what the baseline delivers.
You receive a prioritized remediation roadmap and a posture readout your leadership, owners, and counsel can act on. Where independent attestation is the goal, we prepare the evidence — we never self-certify; formal certification runs through accredited third-party bodies, and the baseline is the on-ramp to ISO/IEC 42001 readiness.
| Week | What happens | From your side |
|---|---|---|
| 1 | Working session; access scoping; AI-feature and data-flow inventory begins. | One technical contact; product walkthroughs. |
| 2–3 | Evidence collection; three-lens scoring; fair-housing control review; cost read. | Short interviews with product, security, and compliance owners. |
| 4 | Readout: posture, prioritized roadmap, remediation plan, attestation path. | A leadership readout session. |
Every step has a human decision gate — nothing is assessed, scored, or reported without your context. Fixed fee, scoped in the first working session once we confirm the AI-system count together.
We bring the baseline instrument and a draft inventory built from your public footprint; you bring the people who know the products. Ninety minutes confirms scope, fee, and timeline — and you leave with our preliminary observations either way.